Elsevier Editorial System(tm) for Energy and
Transcript of Elsevier Editorial System(tm) for Energy and
Citation:Parker, JM and Farmer, D and Fletcher, M (2015) Calibrating whole house thermal models against acoheating test. In: SYSTEM SIMULATION IN BUILDINGS 2014, Liege.
Link to Leeds Beckett Repository record:https://eprints.leedsbeckett.ac.uk/id/eprint/2408/
Document Version:Conference or Workshop Item (Accepted Version)
Creative Commons: Attribution 4.0
The aim of the Leeds Beckett Repository is to provide open access to our research, as required byfunder policies and permitted by publishers and copyright law.
The Leeds Beckett repository holds a wide range of publications, each of which has beenchecked for copyright and the relevant embargo period has been applied by the Research Servicesteam.
We operate on a standard take-down policy. If you are the author or publisher of an outputand you would like it removed from the repository, please contact us and we will investigate on acase-by-case basis.
Each thesis in the repository has been cleared where necessary by the author for third partycopyright. If you would like a thesis to be removed from the repository or believe there is an issuewith copyright, please contact us on [email protected] and we will investigate on acase-by-case basis.
Elsevier Editorial System(tm) for Energy and
Buildings
Manuscript Draft
Manuscript Number: ENB-D-14-00398R3
Title: Obtaining the heat loss coefficient of a dwelling using its
heating system (Integrated coheating)
Article Type: Full Length Article
Keywords: Coheating; Heat loss coefficient; Thermal performance; Energy
signature; Heating system; Performance gap; Heat meter; Energy
efficiency; Building fabric; Whole house heat loss
Corresponding Author: Mr. David Farmer,
Corresponding Author's Institution: Leeds Beckett University
First Author: David Farmer
Order of Authors: David Farmer; David Johnston; Dominic Miles-Shenton
Abstract: This paper presents the methodology, along with some of the
initial findings and observations from tests performed on two dwellings,
of differing construction and form, in which a coheating test was
performed using the dwelling's central heating system; this method is
referred to as integrated coheating. Data obtained during the integrated
coheating tests using a dwelling's heating system have been compared with
data obtained during electric coheating of the same dwelling. In one
instance, integrated coheating test data from one dwelling was compared
to a similar adjoining control dwelling that was simultaneously subject
to an electric coheating test. The results show a good agreement between
the heat loss coefficients (HLC) obtained using a dwelling's own heating
system and those obtained through electrical coheating. Initial analysis
suggests the HLC estimate obtained from integrated coheating is likely to
be more representative of how a dwelling performs in-use. The findings
question the appropriateness of comparing current steady-state HLC
predictions to those derived from in-use monitoring data. Integrated
coheating has the potential to provide a more cost-effective and
informative indication of whole house heat loss than electric coheating,
as it enables in situ quantification of both fabric and heating system
performance.
David Farmer
Centre for the Built Environment
Leeds Sustainability Institute
Leeds Beckett University
BPA223 Broadcasting Place
Woodhouse Lane
Leeds
LS2 9EN
+44 (0)113 812 1938
8th
November 2015
Professor Matheos Santamouris
Editor
Energy and Buildings
Dear Professor Santamouris,
I am writing to submit the revised manuscript entitled, “Obtaining the heat loss coefficient of a dwelling using
its heating system (integrated coheating)”, for consideration for publication in Energy and Buildings.
Please accept my apology for the delay in revising the manuscript, unfortunately I had to take a year off from
my PhD studies due to family circumstances and was unable to re-submit until now.
Thank you for receiving the revised manuscript and for considering it for review. I appreciate your time and
look forward to your response.
If you have any questions, please do not hesitate to contact me.
Yours sincerely,
David Farmer
Cover Letter & Abstract
Obtaining the heat loss coefficient of a dwelling using its heating system (integrated coheating)
David Farmer a*
, Professor David Johnston a, Dominic Miles-Shenton
a
a Centre for the Built Environment, Leeds Sustainability Institute, Leeds Beckett University (formerly known as
Leeds Metropolitan University), BPA223 Broadcasting Place, Woodhouse Lane, Leeds, LS2 9EN, UK.
* [email protected] +441138121938
Abstract
This paper presents the methodology, along with some of the initial findings and observations from tests
performed on two dwellings, of differing construction and form, in which a coheating test was performed using
the dwelling’s central heating system; this method is referred to as integrated coheating. Data obtained during
the integrated coheating tests using a dwelling’s heating system have been compared with data obtained during
electric coheating of the same dwelling. In one instance, integrated coheating test data from one dwelling was
compared to a similar adjoining control dwelling that was simultaneously subject to an electric coheating test.
The results show a good agreement between the heat loss coefficients (HLC) obtained using a dwelling’s own
heating system and those obtained through electrical coheating. Initial analysis suggests the HLC estimate
obtained from integrated coheating is likely to be more representative of how a dwelling performs in-use. The
findings question the appropriateness of comparing current steady-state HLC predictions to those derived from
in-use monitoring data. Integrated coheating has the potential to provide a more cost-effective and informative
indication of whole house heat loss than electric coheating, as it enables in situ quantification of both fabric and
heating system performance.
Keywords
Coheating; Heat loss coefficient; Thermal performance; Energy signature; Heating system; Performance gap;
Heat meter; Energy efficiency; Building fabric; Whole house heat loss.
Response to reviewer
The introduction (section 1) has been revised and expanded to reflect recent developments in
coheating testing. From the available literature it is evident that the primary focus of current research
into coheating regards the analysis of test data. The introduction mentions the most pertinent
conclusion of Bauwens and Roels, 2014 paper, Co-heating test: A state-of-the art, which is that
multiple linear regression is the most sensible method of analysing coheating test data.
Section 2 which relates to the data analysis of the experiments refers to Bauwens and Roels
conclusion that solar aperture should be obtained using multiple linear regression rather than
calculation methods based on glazed areas such as SAP.
I believe I have selected the most relevant conclusions of Bauwens and Roels 2014 paper in regard to
the work presented, which is primarily concerned with an alternative method of providing and
measuring the heat input in a coheating test, rather analysis of the test data. The analysis in the
submitted paper was completed prior to Bauwens and Roels 2014 paper was published; fortunately I
had analysed my data using multiple linear regression.
Detailed Response to Reviewers
1 Introduction
EU [1] and UK [2] regulations are progressively increasing the building fabric energy efficiency standard of
new and existing dwellings driven by the requirement to reduce CO2 emissions and the increasing cost of the
energy required to heat dwellings. A body of evidence is being amassed highlighting a discrepancy between the
predicted and as-built thermal performance of the building fabric which threatens to reduce the desired impact
of these regulatory measures (Stafford et al. [3] and Johnston et al. [4]). This underperformance is commonly
referred to as the ‘performance gap’. In order for the thermal performance of buildings to be quantified a metric
is required: the heat loss coefficient (HLC) is one such metric. The HLC is the rate of heat loss in Watts from
the entire thermal envelope of a building per Kelvin of temperature differential between the internal and external
environments (ΔT) and is expressed in W/K. Obtaining an estimate of a building’s HLC in situ enables a
comparison to be made between the realised performance and predicted performance and enables feedback to
the occupier, building management system and to other stakeholders regarding the thermal performance of the
dwelling.
Comparable metrics to the HLC can be obtained from in-use monitoring data using linear regression based
energy signature analysis techniques (Hammarsten [5], Sjogren et al. [6]). As many of these models rely on
assumptions regarding occupant behaviour, their accuracy must be questioned. Complex dynamic statistical
models are also being identified which aim to isolate the effect of occupant behaviour, enabling identification of
the HLC and other parameters from in-use monitoring data (Bacher and Madsen [7]). Although these methods
could enable a HLC to be isolated from smart metering of an occupied dwelling; verification of their output
parameters against measured baseline values is required to establish their reliability.
The uncertainties associated with occupant behaviour when estimating the HLC in situ can be removed by
physical measurement of an unoccupied dwelling. Physical measurement techniques can be separated into two
distinct categories: disaggregate and aggregate. To estimate the HLC of a building using disaggregate
techniques, the U-value of all thermal elements must be measured (commonly using heat flux plates), along with
the background ventilation rate of the building (pressurisation testing or tracer gas methods), and linear thermal
bridging (Taylor et al. [8]). Estimating the HLC using a combination of disaggregate methods has the advantage
of providing multiple parameters relating to the building fabric which can potentially isolate the cause of any
potential performance gap. However the veracity of the HLC estimate is questionable as it is difficult to ensure
that the U-values measured in situ are representative of the entire element (especially ground floors and bridging
*ManuscriptClick here to view linked References
layers), and measurement of linear thermal bridging is highly complex in a dynamic environment. Although
aggregate methods yield less information regarding individual parameters of the building envelope, they capture
the thermal bridging component of the HLC and can obtain an estimate of the HLC with a lower level of
complexity; one such method is electric coheating.
Electric coheating is a recognised test method for obtaining an estimate of the in situ HLC of a building. A
coheating test involves heating the internal environment of a building to an elevated, homogenous, and constant
temperature with electric resistance heaters and then maintaining that temperature over a number of days
(typically 10-21 days). The power input to the dwelling, as well as the internal and external environmental
conditions, is monitored throughout the test.
Electric coheating has existed in various forms since the late 1970s. It was originally performed overnight as a
test to measure the efficiency of heating systems that cannot be measured directly, such as fireplaces and
furnaces (Sonderegger and Modera [9] and Sonderegger et al. [10]). These early tests found that heating a
building solely with electric resistance heaters meant that the building’s HLC could also be measured. Future
development of the coheating test in the 1980s in the UK (Siviour [11], Everett et al. [12] and Everett [13])
focused upon measurement of the HLC. These works increased the length and complexity of the test and
analysis to accommodate for the dynamic external environment in which coheating tests take place.
The use of coheating increased in the UK following its uptake and development by Leeds Metropolitan
University (now known as Leeds Beckett University); notably during the Stamford Brook Project. Coheating
tests during the Stamford Brook Project identified a substantial performance gap in new dwellings, and helped
quantify the party wall bypass heat loss mechanism (Lowe et al. [14]). Following the Stamford Brook Project,
the electric coheating test method was further refined and developed by Leeds Metropolitan University,
resulting in the 2010 version of LeedsMet’s Whole House Heat Loss Test Method (Wingfield et al. [15]). This
version became recognised as an established test method in the UK when it was incorporated within the Post
Construction and Initial Occupation studies undertaken under the Technology Strategy Boards (now Innovate
UK’s) Building Performance Evaluation Programme [16]. The 2010 version of the test method was significantly
revised in 2013 (Johnston et al. [17]).
In recent years, research efforts have begun to concentrate on the methods used to analyse the data obtained
from coheating tests, rather than the experimental setup or the testing methodology. Work has been undertaken
on the reliability of the coheating test (Bauwens et al. [18]) and on the uncertainties associated with associated
with thermal mass and solar radiation present during a coheating test (Stamp et al. [19]). Most recently, a state-
of-the-art review of the coheating test and the methods used to analyse test data proposes that the most sensible
analysis method to adopt is multiple linear regression (Bauwens and Roels [20]).
Unlike the fan pressurisation test method, which is used to establish the air permeability of dwellings, the
coheating test has not been widely-adopted as a procedure for either regulatory compliance or quality control
purposes. Instead, it remains the preserve of a few academic institutions and specialist consultancy services
(Zero Carbon Hub [21]). There are numerous reasons why the coheating test has seen limited application, which
include: reluctance from the construction industry to acknowledge and research the performance gap; criticism
regarding the precision and accuracy of the coheating test (Butler and Dengle [22]); the duration of the test with
no guarantee of obtaining a confident estimate of the HLC in the limited time available; the test’s restriction to
the heating season (October – March in the UK); the lack of a recognised, standardised test and analysis method;
a lack of experienced testers; and, the time and financial costs associated with undertaking the test (Taylor et al.
[8]).
The financial cost of a coheating test can be disaggregated into costs associated with the time for which a
dwelling must remain unoccupied, as well as the personnel, equipment and energy costs. Dynamic whole house
heat loss test methods exist that are far shorter in duration than the coheating test. These include: ISABELE
(Bouchié et al. [23]), the Quick U-value of Buildings (QUB) method (Mangematin et al. [24]) and the Primary
and Secondary Terms-Analysis and Renormalization (PSTAR) method (Subbaro [25], Subbaro et al. [26]).
However, the HLC estimation obtained by PSTAR has been questioned (Palmer et al. [27]) and the robustness
of methods currently under development (the QUB and the ISABELE method) is yet to be established. Other
financial costs could be significantly reduced by substituting the dwelling’s heating system for the portable
electric heaters that are commonly used to provide the heat input during whole house heat loss tests.
This paper provides details of the methodology and early analysis of the data obtained from experiments
performed on two dwellings where coheating tests were undertaken using the dwelling’s own hydronic central
heating system to provide heat input; a method referred to as integrated coheating. It also compares the results
obtained to the same dwellings undergoing electric coheating in accordance with LeedsMet’s 2010 coheating
method (referred to henceforth as LeedsMet coheating).
2 Estimation of the HLC from a coheating test
Following an initial period during which the building fabric reaches thermal capacitance, a coheating test
assumes the following whole house energy balance (adapted from Siviour [11]):
Q + R.S = (ΣU.A + Cv).ΔT (1)
Where: Q is the total measured power input from space heating (W), R is the solar aperture of the house (m2), S
is the Solar irradiance (W/m2), ΣU.A is the total fabric transmission heat loss (W), Cv is the background
ventilation heat loss (W) and ΔT is the temperature difference between the internal and external environment.
The whole house energy balance equation can be rearranged to show that:
HLC = (Q + R.S) / ΔT (2)
The HLC is typically estimated using a linear regression-based quasi-steady-state analysis of the data obtained
during the test period. The raw HLC can be obtained from the slope of a simple linear regression analysis in
which Q is the dependent variable and ΔT is the independent variable, though this does not account for the
effect of solar radiation. The power provided by solar radiation to the dwelling during a coheating test (R.S) is
not measured directly, rather its effect is observed in a measured reduction in the power required to maintain a
constant internal temperature, which is manifested in a reduction in the raw HLC. Solar radiation not only
results in solar transmittance through glazed elements which is absorbed within the building, but it is also
absorbed by the opaque exterior surfaces of the building, causing a reduction (or temporary net reversal) of heat
flow through external elements. It is therefore not advisable to calculate the solar aperture (R) of a dwelling
based solely on the glazed properties of a dwelling, as in the SAP 2009 methodology (BRE [28]). A more
appropriate method is to introduce solar irradiance into a multiple regression analysis in which ΔT and S are
independent variables and Q the dependent variable (Bauwens and Roels [20]). The multiple regression analysis
produces regression coefficients for ΔT (the HLC) and for S (the solar aperture). It is then possible to correct Q
for R.S, which can be plotted in a simple linear regression with solar corrected power (Q + R.S) as the
dependent variable and ΔT as the independent variable. The resulting slope is the HLC estimate. Wind speed
can also be included in the multiple regression analysis, with the HLC estimate being corrected to include the
wind speed regression coefficient. It is common for the constant to be excluded (regression through origin), as
the model assumes that at zero ΔT with zero solar irradiance and no wind, there is no heat transfer from the
dwelling. All estimates of the HLC in this paper have been obtained using the multiple regression based
methods detailed in this section with the constant omitted.
Although the appropriateness of a simple linear model and the application of steady-state analysis to a dynamic
system has been questioned (Baker and van Dijk [29]), in practice the coheating tests are undertaken over a
sufficient length of time to minimise or smooth out some of the dynamic effects, such as thermal storage and
inertia.
Data used in the regression analysis comprise 24 hour mean values for each variable to account for the diurnal
cycle. A 6 a.m. – 6 a.m. 24 hour interval is often used as this provides the greatest opportunity for daytime solar
radiation absorbed by the building fabric to be released within each period, thus reducing errors attributable to
solar thermal storage from one 24 hour period influencing the results from another and increasing the reliability
of the regression analysis.
3 Method
In order to establish whether the integrated method can produce a useful estimate of the HLC, a confident
baseline estimate of the in situ HLC of each test dwelling is required. In the absence of an international standard
for coheating, the LeedsMet coheating method was selected for the electric coheating tests. HLC estimates
obtained using the integrated method can then be validated against the baseline HLC estimate.
3.1 Comparison of LeedsMet and integrated coheating test methods
Table 2 lists the principal considerations when conducting a LeedsMet coheating test and compares these
against the approach taken in experiments undertaken using the integrated method.
Table 2: Comparison of the LeedsMet and integrated coheating methods
From Table 2, it can be seen that the notable differences between the two approaches relates to the provision and
distribution of heat within the dwelling and the measurement of power input from space heating (Q); these will
be explored in greater detail.
The significant difference in the two methods is the use of the dwelling’s own heating system to provide space
heating in lieu of the electric resistance heaters. The integrated method shares similarities with a coheating test
undertaken on a school, in which for practical reasons, a proportion of the building was heated using the heating
system, with other areas heated electrically (Zabot et al. [30]). Electric resistance heaters are used in the
LeedsMet coheating tests because they are effectively 100% efficient at the point of use; this enables a reliable
measure of Q to be obtained using an energy meter. Portable electric resistance heaters can be strategically
placed throughout the dwelling to ensure heat input into all areas of the dwelling.
The efficiency of modern condensing gas boilers is in the region of 90%; operational efficiency can vary
depending upon the boiler load and climate conditions. It is therefore not possible to obtain a reliable measure of
Q using gas metering alone and this has prevented its use in coheating tests. However, the heat output of gas
boilers can be measured using a heat meter, and these are becoming more affordable and prevalent. A heat meter
comprises a flow sensor, a pair of temperature sensors, and an integrator which calculates the heat dissipated
into the dwelling using the heat transmission formula:
Q = ρ.V.cp.(tf – tr) (3)
Where: Q is the quantity of heat given up or absorbed (W), ρ is the density of the heat transfer medium (kg/m3),
V is the volume flow rate (m3/s), cp is the specific heat capacity at constant pressure (kJ/kgK), tf is the
temperature of the liquid in the flow pipework (K) and tr is the temperature of the liquid in the return pipework
(K).
A shortcoming of measuring Q with a heat meter is that measurement error is greatest at lower flow rates, which
creates more uncertainty about the value of Q at smaller temperature differentials. In addition, heat meters do
not account for the additional heat input from boiler casings housed within the thermal envelope (though this is
likely to be a small proportion of total heat input). Heat metering could provide a better measurement of Q in
dwellings where the heat input was metered on entry to the dwelling, such as where a district or communal
heating system is used, or some heat pump configurations. In situations where the plant used to generate heat is
contained within the thermal envelope, the value of Q is more likely to be:
Q = Qhm + Qp (4)
Where: Q is the total measured power input from space heating (W), Qhm is the heat input to the dwelling
measured by the heat meter (W), Qp is the heat gain from the heat generation plant (W).
During LeedsMet coheating tests a relatively homogenous air temperature throughout the test dwelling is
facilitated by the use of air circulation fans, ensuring a similar ΔT throughout the building envelope. The
integrated method of coheating relies upon natural convection from the heating system radiators to obtain a
homogenous temperature throughout the dwelling. However, some stratification within a test dwelling will
inevitably occur and some areas of a test dwelling might not contain any, or sufficient, heat emitters to produce
a homogenous internal temperature. If the test dwelling has a complex internal arrangement, this may also
hinder heat distribution throughout the internal environment. A lack of homogeneity in internal temperature will
complicate the calculation of the mean internal temperature and could also result in some areas of the building
fabric experiencing differing patterns of thermal storage and release than others.
The internal temperature during LeedsMet coheating is controlled by thermostatic temperature controls
connected to each heater, which facilitates a homogenous internal environment. The modern fuzzy logic
thermostatic temperature controllers increasingly used in coheating maintain an almost constant internal
temperature, which minimises dynamic fluctuations within the building fabric emanating from the internal
environment. As the energy balance is based upon steady state assumptions, this is an important factor.
The level of internal temperature control during integrated coheating will largely be dictated by the control
system of the test dwelling. In the UK, it is common practice to have a wall mounted thermostat which controls
the flow of heat from the boiler to all radiators within a particular heating zone. Heat delivery to each zone is
dictated by this thermostat and TRVs are usually fitted on all radiators within each zone, except for the radiator
closest to the control thermostat. The low accuracy and sensitivity of TRVs, compared to the control thermostat,
means there will be less control of temperature in rooms where these are fitted. In addition, no heat will be
delivered to these radiators if the thermostat controlling flow from the boiler is satisfied. More advanced digital
zoning and radiator control valves should increase the level of temperature control within a dwelling.
3.2 Test dwellings
Obtaining empty test dwellings is consistently a problem in field studies of this nature. For these experiments
three test dwellings were made available by Joseph Rowntree Housing Trust (JRHT). Although two of the
dwellings are of a common construction type found in the UK, the sample size is small and the dwellings are in
a similar geographical area, therefore the sample cannot be considered truly representative of the UK housing
stock. The test dwellings are illustrated in Fig. 1 and a summary of their characteristics is provided in Table 3.
Fig. 1: Test dwelling A (left) and test dwellings B and C (right)
Table 3: Test dwelling characteristics
3.3 Experimental procedure
Each test dwelling was subject to three tests:
1. LeedsMet coheating to obtain a baseline HLC estimate.
2. Integrated coheating using gas-fired central heating system with air circulation fans.
3. Integrated coheating using gas-fired central heating system only.
The rationale for performing the integrated coheating test with the use of air circulation fans is that it could yield
further understanding of the effects of air circulation during coheating tests, as well as potentially isolating
differences caused by the variation in the measurement of Q between coheating methods.
A party wall separates test Dwellings B and C. Dwelling C was subject to LeedsMet coheating throughout the
experimentation period and acted as a control dwelling which enables the response of similar dwellings using
different coheating methods to be compared. Any anomalies in the estimate of the HLC that might be caused by
the external environmental conditions, by inherent problems with the analysis method, or by unidentified factors
are therefore not attributable to the variation in the test method. Maintaining the internal temperature of
Dwelling C isothermal to Dwelling B also minimises heat transfer between the party elements of these
dwellings.
The set-point temperature for the fuzzy logic thermostatic temperature controllers in the LeedsMet tests was
25oC. The same set-point temperature was selected for the integrated test, which involved setting the control
thermostat(s) to 25oC and the (TRVs) on each radiator to a setting which the manufacturers’ data sheet
suggested would maintain a local temperature of 25oC. The programmer for the central heating system was set
to manual override (always on) so that boiler output was controlled by the zone thermostats only.
Prior to the experiment, the flow and return pipework from the boiler of Dwelling B was fitted with a heat
meter. Dwelling A had previously been fitted with a heat meter as part of a previous in-use monitoring project.
Both the heat meters used were Sontex Supercal 539 (uncertainty ≤ ± 5%) that had a pulse output registering
one pulse per 100 Wh of energy delivered.
Energy consumption, along with internal and external environmental data in all tests, was logged at ten minute
intervals throughout the experiments using an Eltek Squirrel RX250AL data logger. Missing data was corrected
using linear interpolation. The gas consumption of Dwelling A was also logged at ten minute intervals.
Electrical energy consumption was measured using an using an Elster A100C energy meter (uncertainty ± 1%)
with pulse output, registering one pulse per 1 Wh. Internal temperatures were measured in each room using
shielded RTD temperature sensors (uncertainty ± 0.1 K), whilst the external environmental conditions were
measured using a Vaisala WXT520 weather station containing a thermistor (uncertainty ± 0.3 K). Irradiance was
measured using a vertically mounted south-facing Kipp and Zonen CMP 11 pyranometer. The internal and
external environmental monitoring systems remained the same between experiments to reduce the number of
variables changing between tests.
During the experiment measurements of heat flux were undertaken using Hukseflux HFP01 heat flux plates in
accordance with ISO 9869:1994 [31] to provide greater insight into the behaviour of the dwellings between
tests. In situ U-values were calculated using the averaging method contained within ISO 9869.
4 Results and discussion
4.1 HLC estimates
4.1.1 Dwelling A
Fig. 2 illustrates the linear regression-based solar corrected HLC estimation of Dwelling A from the coheating
test scenarios. HLCs from all test scenarios are shown in Table 4.The baseline HLC estimate of 133.7 W/K
differs by <1% from the HLC estimate of 132.9 W/K obtained from a previous coheating test undertaken on this
dwelling in 2010, almost 3 years earlier (Miles-Shenton et al. [32]). This increases confidence in the baseline
value and suggests that LeedsMet coheating may be reasonably precise.
Fig. 2: Linear regression estimation of Dwelling A HLC from each coheating test scenario
Table 4: Summary of the estimations of the HLC of Dwelling A obtained from the three test scenarios
A one-way ANOVA was used to test for differences in the HLC estimates obtained from the three tests. Data
comprises 24 hour solar corrected HLCs. There was a statistically significant difference between the three tests,
F (2,97) = 52.79 p = <0.001. Gabriel post-hoc comparisons of the three groups indicate that the integrated test
(M = 117.3, 95% CI [114.5, 120]) resulted in a significantly lower estimate of the HLC than the integrated test
using air circulation fans (M = 135.9, 95% CI [132.4, 139.3]), p = <0.001 and baseline test (M = 134.8, 95% CI
[131.9, 137.7]), p=<0001. The analysis also revealed that there was no statistically significant difference
between the baseline test and integrated with fans test (p = 0.946).
The HLC estimate obtained from the integrated test is below the predicted HLC of 120.1 W/K, however the
LeedsMet and integrated with fans tests both produced an estimate greater than the predicted HLC. Some of the
decrease in HLC estimate obtained from the integrated method could be attributed to a reduction in convective
heat transfer resulting from the absence of fans. However, it is thought the internal arrangement of the dwelling
on the second floor prevented some areas of the thermal envelope from reaching the same temperature as the
living areas (principally the loft and the space behind the knee walls on the second floor, all of which were
contained within the thermal envelope – hatches to these areas remained open during all tests). Thermography
and temperature measurements undertaken during the integrated coheating test revealed these spaces to be
approximately 1-2 K cooler than adjacent rooms. These spaces were primarily heated by TRV controlled
radiators in adjacent rooms which ceased providing heat input once the room air temperature reached the set-
point. This resulted in under heating of the dwelling, which was not accounted for in the ΔT calculation. The
limited air exchange between these areas and the living areas effectively reduced the size of the thermal
envelope, as they were providing additional thermal resistance between the external environment and heated
living areas.
During the LeedsMet coheating test, a thermostatically controlled portable electric heater and fan was placed in
the loft, which maintained a constant and homogenous temperature in this space. In addition, air circulation fans
were used to provide heated air to the knee wall voids. The location of all circulation fans was not changed for
the integrated coheating with fans test, although a slight adjustment was made to the master bedroom fan
position to direct air into the loft space. The provision of heat to all areas of the thermal envelope may be
responsible for the higher HLC estimates obtained during the LeedsMet coheating test.
4.1.2 Dwellings B and C
4.1.2.1 Dwelling B
Fig. 3 illustrates the linear regression-based solar and wind corrected HLC estimation of Dwelling B from the
coheating test scenarios, HLCs from all test scenarios are shown in Table 5. The largest discrepancy occurs
between the baseline test HLC estimate and the integrated (fans) test estimate.
Fig. 3: Linear regression estimation of Dwelling B HLC from each coheating test scenario
Table 5: Summary of the estimations of the HLC of Dwelling B obtained from the three test scenarios
A one-way ANOVA was used to test for differences in the HLC estimates obtained from the three tests. Data
comprises 24 hour corrected HLCs. There was a statistically significant difference between the three tests, F
(2,49) = 22.866 p = <0.001. Gabriel post-hoc comparisons of the three groups indicate that the baseline test (M
= 114.4, 95% CI [112.3, 116.5]) resulted in a significantly higher estimate of the HLC than the integrated test
using air circulation fans (M = 104.1, 95% CI [101.5, 106.8]), p = <0.001 and integrated test (M = 105, 95% CI
[101.7, 108.2]), p=<0001. There was no statistically significant difference between either of the integrated tests
(p = 0.96).
The use of circulation fans appeared to have no statistically significant impact on the HLC estimate for the
dwelling. The comparatively low impact of circulations fans in Dwelling B compared with Dwelling A could be
explained by the regular internal arrangement of Dwelling B and the provision of heat to all areas within the
thermal envelope, aided by a multi-zone central heating system providing control of the internal temperature.
4.1.2.2 Dwelling C (control)
Table 6 provides the solar and wind corrected HLC estimations of control Dwelling C during the period in
which the baseline test and integrated coheating tests were being undertaken in Dwelling B.
Table 6: Summary of control Dwelling C HLC estimations obtained using LeedsMet coheating
A one-way ANOVA was used to test for differences in the HLC estimates obtained from LeedsMet coheating of
the control dwelling during the three tests. Data comprises 24 hour solar corrected HLCs. There was a
statistically significant difference between the three tests, F (2,49) = 10.68 p = <0.001. Gabriel post-hoc
comparisons of the three tests indicate that the integrated test using air circulation fans (M = 98.9, 95% CI [96.6,
101.2]) resulted in a significantly lower estimate of the HLC than the baseline test (M = 105.5, 95% CI [103.5,
107.6]), p = <0.001 and integrated test period (M = 102.8, 95% CI [100.5, 105.1]), p=0.042. There was no
statistically significant difference between the baseline test and integrated test (p = 0.197).
4.1.2.3 Comparison between test Dwelling B with control Dwelling C
From Fig. 4 it can be seen that both dwellings’ 24 hour HLC estimates tend to respond in similar fashion to
dynamic factors present during the baseline LeedsMet coheating test (R2 = 0.816). The outlier near the top right
of the regression line represents the 24 hour period following the greatest change in external temperature during
the test, which is symptomatic of the steady-state coheating analysis. This strong relationship enables Dwelling
C to be used as a control dwelling with a reasonable degree of confidence. A paired-samples t-test was
conducted to compare the 24 hour HLC estimates of Dwelling B and Dwelling C during the baseline
measurement test. This revealed that there is a significant difference in the HLC estimates for Dwelling B
(M=114.4, SD=4.7) and Dwelling C (M=105.5, SD=4.5); t(40)= 6.259, p = <0.001; thus the control dwelling
will only be used to compare relative change in the HLC estimates.
Fig. 4: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) during the
baseline HLC estimation period
Fig. 5 shows a reasonable correlation (R2 = 0.624) between 24 hour HLC estimates during the integrated test in
which air circulation fans were used. In Fig. 6 it can be seen that the correlation was weaker (R2 = 0.468) during
the integrated test without air circulation fans. The lower correlation evident during the integrated tests could
highlight the comparative weakness of a dwelling’s heating system to respond quickly to changes in ΔT,
compared to the heating equipment installed during a LeedsMet coheating test. This could be due to the lower
thermal inertia of the electric resistance heaters used in LeedsMet coheating compared to the central heating
system. Further analysis is required to establish if this is the case.
Fig. 5: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) obtained
during the integrated test with fans
Fig. 6: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) obtained
during the integrated test
The HLC estimates derived from the integrated tests in Dwelling B were significantly lower than the baseline
estimate. The HLC estimate from LeedsMet coheating of control Dwelling C during the integrated test with fans
was also significantly lower than the baseline estimate.
In situ measurements of heat flux on the external walls of the control and test dwellings showed a reduction in
thermal transmittance from the baseline test period to the integrated (fans) test period, see Table 7. The
reduction in external wall thermal transmittance accounts for 49% of the total reduction in HLC estimate of the
control dwelling and 43% of the test dwelling.
Table 7: Reduction in the thermal transmittance of the external walls of the test dwellings from the baseline test
to the integrated test with fans
The reduction in thermal transmittance is attributed to a continual decrease in the moisture content of the
building fabric resulting from prolonged heating during the testing programme. Both dwellings were continually
heated from October 2012 to April 2013. Hygrothermal simulations have shown that the thermal transmission of
the building fabric reduces during the first years of a dwellings operation as heating drives residual construction
moisture from the fabric (Holm et al. [33]). It is probable that other thermal elements also experienced a
reduction in thermal transmittance. Hence, it is thought that a significant proportion of the discrepancy between
the baseline test and integrated tests HLC estimates can be attributed to a reduction in the thermal transmittance
of the test dwelling. It is therefore probable that if the thermal transmission of the building fabric had remained
stable across the test programme, the discrepancy between the baseline and integrated tests HLC estimations
would have been less.
4.2 Internal temperature control
It can be seen from Fig. 7 and Fig. 8 that the LeedsMet coheating maintains a closer control over internal
temperature compared to the integrated test methods. The temperature remained within a generally stable range
(amplitude 0.1 K) in each zone of the test dwellings, except for periods of high solar gain which are denoted as
outliers.
Fig. 7: Box plot of the mean hourly internal temperature in Dwelling A during the three test scenarios
Fig. 8: Box plot of the mean hourly internal temperature in Dwelling B (test) and Dwelling C (control) during
the three test scenarios
The reason for the lower positive skew and number of outliers during the integrated test scenarios in Dwelling B
compared to Dwelling A is thought to be due to under-heating occurring in the north facing rooms in Dwelling
B during periods of high solar radiation. High solar gains resulted in overheating in the south-facing rooms
containing the wall mounted thermostats; this caused the boiler to cease supplying heat to the rest of the
dwelling. The ΔT between the north and south rooms of Dwelling B meant that the mean internal temperature of
Dwelling B remained close to the thermostat set point.
Individual zone temperatures during integrated coheating were characterised by an oscillating pattern around the
mean temperature, with amplitude of ~ 0.5 K in Dwelling A and 1 K in Dwelling B. The mean internal
temperature in Dwelling B was maintained closer to the thermostat set point of 25oC and remained more stable
and homogenous throughout the integrated test periods than in Dwelling A. The closer control in Dwelling B
was probably due to its zoned heating configuration, radiator distribution and dwelling form. The use of air
circulation fans reduced the interquartile range of the mean internal temperature in both test dwellings. This
suggests that the use of air circulation fans can facilitate greater control of internal temperatures.
The small HLC estimate discrepancy between the integrated test involving fans and the baseline test in Dwelling
A would suggest that the higher amplitude of the oscillation in internal temperatures has a low impact on the
HLC estimate. However, the LeedsMet approach facilitates a stable and homogenous temperature throughout a
test dwelling and ensures greater precision between tests. The experiments undertaken highlight the variation in
heating characteristics that can be experienced between differing dwellings.
4.3 Other findings
The test programme also highlighted issues with the installation of the central heating systems. In Dwelling A, a
TRV was fitted to the hall radiator which was close to the main control thermostat. This resulted in problems
with overheating in other areas of the as the TRV was restricting the supply of heat to the hall causing the boiler
to cycle excessively (causing the loss of 8 days data). This was resolved by setting the hall TRV to maximum. In
all three dwellings the TRVs were incorrectly positioned, vertically at the top of the radiator, which will have
reduced their effectiveness in heating the dwelling. The zone thermostats controlling the boiler interlock in
Dwelling B were wired to the wrong actuator valves, resulting in the ground floor thermostat controlling the
first floor heating and vice versa; this issue was responsible for the hiatus period prior to rectification.
Dwelling A’s boiler efficiency was calculated by dividing the energy output of the boiler measured by the heat
meter (Q) by the energy supplied to the boiler, calculated from the metered gas supply. The calculated boiler
efficiency of 84% during the integrated tests was lower than the boiler’s SEDBUK rating of 90.1%. An
additional advantage of using the dwellings own gas-fired central heating system to undertake an integrated
rather than a LeedsMet coheating test, is a reduction in the energy costs and CO2e emissions. If an 84% efficient
gas boiler is used to provide the heat input during a coheating test, a 54.7% CO2e emission reduction1 and
63.4% cost reduction2 can be achieved compared to mains electric heat input
3.
5 Conclusions
The initial findings show good agreement between the LeedsMet coheating and the integrated method, though
further tests will need to be conducted to enable any firm conclusions to be drawn about the precision or
accuracy of the integrated method. Although the LeedsMet coheating method is likely to be the more precise,
such precision might not be required for basic compliance or quality assurance procedures. Instead, it may be
more appropriate in the academic sphere or in dwellings that have been designed to have a very low HLC, such
as Passivhaus.
The findings presented suggest that using the dwelling’s own heating system to deliver heat during a coheating
test is acceptable and that a heat meter measuring boiler output is a suitable device for obtaining a reasonable
estimate of the power input from the space heating system. The discrepancy between the boiler efficiency stated
by the manufacturer and that measured in situ means that calculating heat input to a dwelling from metered gas
consumption cannot provide an accurate estimate of heat input into a dwelling.
1 Based upon UK 2013 values for natural gas of 0.18404 kgCO2e and mains electricity (including transmission
losses) of 0.48357 kgCO2e (DEFRA [34]). 2 Based upon the mean UK 2012 price per kWh for mains natural gas of £0.0475 and mains electricity of
£0.1544 (DECC [35]). 3 Based upon a mean internal temperature of 25
oC, and the mean 2011 – 2013 heating season (November –
February) external temperature for York, UK of 4.5oC (BizEE [36]).
The effect of introducing air circulation fans in Dwelling A suggests that the movement of air, and thus heat,
within the dwelling could be the greatest source of discrepancy in the HLC estimates observed between the
LeedsMet coheating test and the integrated coheating test. This implies that the integrated coheating method is
more appropriate for use in dwellings without a complex internal arrangement, or large areas without heat
provision. However, the integrated method may produce an estimate of a dwelling’s HLC that is more
representative of how a dwelling performs in-use. The use of portable electric heaters and air circulation fans in
LeedsMet coheating is effective at creating a homogenous internal temperature, which ensures consistency
between tests and a more valid comparison with steady-state HLC predictions. However, it is unlikely that a
dwelling’s heating system and occupant behaviour will replicate these conditions in the field. Therefore it can be
assumed that the HLC derived from an energy signature analysis of an occupied dwelling will be closer to that
obtained from integrated coheating rather than LeedsMet coheating. Thus, in the case of Dwelling A, both the
integrated coheating test and energy signature analysis may fail to identify the fabric performance gap, or may
underestimate the scale of the gap. This brings into the question the legitimacy of comparing a HLC obtained
from in-use monitored data to HLC predictions using current steady-state models. As a consequence, a more
complex HLC prediction model, which accounts for the provision and movement of heat within a dwelling, may
be required to enable the accurate quantification of a fabric performance gap from in-use monitoring data.
The information provided by an integrated coheating test reveals more than just an estimate of the HLC of a
dwelling and fabric performance; it also provides information relating to the performance and characteristics of
the space heating system, as it tests the dwelling as an entire system. Experiments involving the integrated
coheating test identified significant issues with the heating systems in the test dwellings. An integrated
coheating test performed after dwelling completion could be undertaken as part of a holistic whole house
commissioning process. The procedure could be used to identify issues relating to the heating system, such as
inappropriate delivery and control of heat and could highlight issues of space heating system efficiency. Thus,
integrated coheating has the potential to identify fabric and heating system performance gaps.
The reduction in energy cost and labour required to transport and distribute coheating equipment means that
integrated coheating should be a more financially viable whole house heat loss test than electric coheating.
However, the period required to perform the test could be the greatest barrier preventing its use as a widely used
regulatory compliance tool. Conducting an integrated coheating test in combination with a dynamic test to
reveal characteristics of a building which relate to thermal mass properties could provide useful parameters
which inform the building management system as to the behaviour of the whole house as a system, as well as
assisting with the identification of a dwelling’s energy signature from statistical analysis of future in-use data.
6 Acknowledgements
The authors would like to thank Professor Chris Gorse for backing and resourcing this project. Matthew Brooke-
Peat for the assistance he provided with some of the fieldwork and John Bradley for proof-reading. Thank you
to Owen Daggett along with the Joseph Rowntree Housing Trust for providing the test dwellings and for their
long-standing support for the Centre for the Built Environment (CeBE) research group.
7 References
[1] Energy Performance Building Directive, 2010/31/EU
[2] HM Government. (2010) Approved Document L1A: Conservation of Fuel and Power in New Dwellings. The
Building Regulations 2010. London, NBS.
[3] Stafford, A., Bell, M. and Gorse, C. (2012) Building Confidence - a working paper. Centre for Low Carbon
Futures report No. 008. Available from: <www.lowcarbonfutures.org.uk/reports/research-reports> [Accessed
January 2014].
[4] Johnston, D. Miles-Shenton, D. and Farmer, D. (2015) Quantifying the Domestic Building Fabric
‘Performance Gap’. Building Services Engineering Research & Technology (BSER&T). Volume 36, No.5,
September 2015, pp.614–627.
[5] Hammarsten, S. (1987) A critical appraisal of energy-signature models. Applied Energy, 26 (2) pp. 97-110.
[6] Sjogren, J-U., Addersson, S. and Olofsson, T. (2009) Sensitivity of the total heat loss coefficient determined
by the energy signature approach to different time periods and gained energy. Energy and Buildings, 41 pp.
801-808.
[7] Bacher, P. and Madsen, H. (2011) Identifying suitable models for the heat dynamics of buildings. Energy and
Buildings, 43 (7) pp. 1511-1522.
[8] Taylor, T., Counsell, J and Gill, S. (2013) Energy efficiency is more than skin deep: Improving construction
quality control in new-build housing using thermography. Energy and Buildings, 66 pp. 222-231.
[9] Sonderegger, R. C. and Modera, M. P. (1979) Electric co-heating: A method for evaluating seasonal heating
efficiencies and heat loss rates in dwellings. In: Proceedings of the Second International CIB Symposium,
Energy Conservation in the Built Environment, 28th
May – 1st June, 1979, Copenhagen. LBL-8949.
[10] Sonderegger, R. C., Condon, P. E. and Modera, M. P. (1980) In-situ measurements of residential energy
performance using electric co-heating. ASHRAE Transactions, 86 (I), LBL-10117.
[11] Siviour, J. (1981) Experimental Thermal Calibration of Houses. Technical report for the Electricity
Council Research Centre, Chester, United Kingdom.
[12] Everett, R. Horton, A. and Doggart, J. (1985). Linford Low Energy Houses. Milton Keynes, Energy
Research Group, Open University.
[13] Everett, R. (1985) Rapid Thermal Calibration of Houses. Technical Report ERG 055 for the Science and
Engineering Research Council, Milton Keynes, UK.
[14] Lowe, R., Wingfield, J. Bell, M. and Bell, J. (2007) Evidence for heat losses via party wall cavities in
masonry construction. Building Services Engineering Research and Technology, 28 (2) pp. 161-181.
[15] Wingfield, J., Johnston, D., Miles-Shenton, D. & Bell, M. (2010) Whole House Heat Loss Test Method
(Coheating), Leeds Metropolitan University. Available from:
<http://www.leedsbeckett.ac.uk/as/cebe/projects/coheating_test_protocol.pdf> [Accessed January 2014].
[16] Technology Strategy Board (2012) Building Performance Evaluation. London, Technology Strategy Board.
Available from:
<http://webarchive.nationalarchives.gov.uk/20130221185318/www.innovateuk.org/content/competition/buildin
g-performance-evaluation-.ashx> [Accessed February 2014].
[17] Johnston, D., Miles-Shenton, D., Farmer, D. and Wingfield, J. (2013). Whole House Heat Loss Test Method
(Coheating). Leeds Metropolitan University, June 2013. Available from:
<http://www.leedsbeckett.ac.uk/as/cebe/#> [Accessed January 2014].
[18] Bauwens, G. Standaert, P. Delcuve, F. and Roels, S. (2012) Reliabilty of co-heating measurements. In
Proceedings of BSO12: IBPSA-England First Building Simulation and Optimisation Conference 2012 (BSO12),
10-11th
September 2012, Loughborough, UK, pp. 348-355.
[19] Stamp S Lowe R and Altamirano-Medina, H An Investigation into the Role of Thermal Mass on the
Accuracy of Co-heating Tests Through Simulations & Field Results. In Proceedings of BS2013: 13th
Conference of International Building Performance Simulation Association (IPBSA), Building Simulation 2013,
August 25-28, Chambery, France, pp. 39-46.
[20] Bauwens, G. and Roels, S. (2014) Co-heating test: A state-of-the-art. Energy and Buildings, 82 pp. 163-
172.
[21] Zero Carbon Hub. Carbon Compliance for Tomorrow’s New Homes: a review of the modelling tool and
assumptions. Topic 4: Closing the gap between designed and built performance. August 2010. London, Zero
Carbon Hub, London. Available from:
<http://www.zerocarbonhub.org/sites/default/files/resources/reports/Carbon_Compliance_Topic%204_Closing_
the_Gap_Between_DvAB.pdf> [Accessed February 2014].
[22] Butler, D. and Dengle, A. (2013) Review of co-heating test methodologies. Report F54. Milton Keynes,
NHBC Foundation.
[23] Bouchié, R. Alzetto, F. Brun, A. Boisson, P. and Thebault, S. (2014) Short Methodologies for In-situ
Assessment of the Intrinsic Thermal Performance of the Building Envelope, In Sustainable Places 2014, Nice.
[24] Mangematin, E., Pandraud, G. and Roux, D. (2012) Quick Measurements of Energy Efficiency of Buildings.
C. R. Physique 13 pp. 383-390.
[25] Subbarao, K. (1988) PSTAR - Primary and Secondary Terms-Analysis and Renormalization: A Unified
Approach to Building and Energy Simulations and Short-Term Testing – A Summary. September 1988.
SERI/TR-254-3347. Colorado, USA, Solar Energy Research Institute.
[26] Subbarao, K., Burch, J. D., Hancock, C. E., Lekov, A. and Balcomb, J. D. (1988) Short-Term Energy
Monitoring (STEM): Application of the PSTAR Method to a Residence in Fredericksburg, Virginia. TR-3356.
Colorado, USA, Solar Energy Research Institute.
[27] Palmer, J., Pane, G., Bell, M. and Wingfield, J. (2011) Comparing primary and secondary terms analysis
and re-normalisation (PSTAR) test and co-heating test results. A Report to the Department for Communities
and Local Government, HMSO, London. ISBN:978 1 4098 2902 7.
[28] BRE (2009) The Government’s Standard Assessment Procedure for Energy Rating of Dwellings – 2009
Edition – v9.9, BRE, Garston, Watford
[29] Baker P.H., van Dijk H.A.L. (2008). PASLINK and dynamic outdoor testing of building components,
Building and Environment, 43 (2) pp. 143-151.
[30] Zabot, S., Mahajan, S., Bhargava, A., Soma, L and Fumagalli, S. (1987) A low-cost methodology for
thermal performance monitoring of public schools in the Lombardy Region. Energy and Buildings, 10 (1) pp.
59-71.
[31] ISO (1994) ISO 9869:1994 Thermal insulation – Building elements - In situ measurement of thermal
resistance and thermal transmittance. Geneva, Switzerland, International Organization for Standardisation.
[32] Miles-Shenton, D. Wingfield, J. Sutton, R. and Bell, M. (2010) Temple Avenue Project Part 1: Temple
Avenue Field Trial – Evaluation of Design & Construction Process and Measurement of Fabric Performance of
New Build Dwellings. Leeds, UK, Leeds Metropolitan University. Available from:
<http://www.leedsbeckett.ac.uk/as/cebe/projects/tap/> [Accessed February 2014].
[33] Holm, A., Sedlbauer, J., Radon, J. and Künzel, H.M. (2002). Influence of building moisture on the
hygrothermal performance of buildings. IPB-Report 398. Munich, Fraunhofer.
[34] DEFRA (2013) Government conversion factors for company reporting. London, Department for
Environment Food & Rural Affairs. Available from:
<http://www.ukconversionfactorscarbonsmart.co.uk/LandingPage.aspx> [Accessed February 2014].
[35] DECC (2013) Quarterly Energy Prices. December 2013. London, Department for Energy and Climate
Change. Available from: < https://www.gov.uk/government/collections/quarterly-energy-prices> [Accessed
February 2014].
[36] BizEE (2013) Degree Days.net - Custom Degree Day Data [Internet], BizEE Software Limited. Available
from: <http://www.degreedays.net/> [Accessed February 2014].
Tables and captions
Table 1: Nomenclature
Term Symbol Unit
Heat loss coefficient HLC W/K
Indoor-outdoor air temperature difference ΔT K
Total measured power input from space heating Q W
Solar aperture coefficient R m2
Solar irradiance S W/m2
Total transmission heat loss ΣU.A W
Background ventilation heat loss Cv W
Density of the heat transfer medium ρ kg/m3
Volume flow rate V m3/s
Specific heat capacity at constant pressure cp kJ/kgK
Temperature of the liquid in the flow pipework tf K
Temperature of the liquid in the return pipework tr K
Heat input to the dwelling measured by the heat meter Qhm W
Heat gain from the heat generation plant Qp W
Table(s) with Caption(s)
Table 2: Comparison of the LeedsMet and integrated coheating methods
LeedsMet coheating Integrated method
Internal space heating Portable electrical direct
resistance heaters (100%
efficient as all electrical
energy consumed is ultimately
liberated as heat)
Dwelling’s own installed heating system. Typically in
the UK, hydronic gas central heating system with
condensing gas boiler (~90% efficient) feeding
radiators as the heat emitters. District heating, electric
heaters and heat pumps are also suitable
Power measurement of
internal space heating
Energy meters (uncertainty +/-
1%)
Heat metering of boiler output (uncertainty +/- 5%),
gas meter with correction applied, heat metering of
district heating or heat pumps at point of entry to the
thermal envelope. Minimum heat meter resolution of
100 Wh recommended.
Homogenised internal
temperature
Use of air circulation fans Natural convection
Internal temperature
control
Thermostatic temperature
controller for each heater
Dwelling’s heating control. Typically in the UK, wall
mounted thermostat(s) and thermostatic radiator valves
(TRVs)
Internal temperature
measurement
Temperature sensors in each
zone
Temperature sensors in each zone. Potentially, a
building management system
External
environmental
conditions
External weather station with
vertical south facing
pyranometer
External weather station with vertical south facing
pyranometer or local weather station data
Table 3: Test dwelling characteristics
Dwelling A Dwellings B and C
Heating system Hydronic central heating system with
radiators as heat emitters. 24 kW gas-
fired condensing gas boiler
Hydronic central heating system with
radiators as heat emitters. 24 kW gas-
fired condensing gas boiler
Heating controls One zone with bimetallic thermostat in
the hall on the ground floor. TRVs on
all radiators except bathroom
Two zone, with bimetallic thermostats
in living room on the ground floor and
main bedroom on the first floor. TRVs
on radiators in all other rooms
Build form Detached. 2 storey plus room in roof.
South facing winter garden
Semi-detached. 2 storey
Internal arrangement Irregular Regular
Total floor area 155 m2
90 m2
Dwelling volume 411 m3 225 m
3
External wall area 186 m2
84 m2
Glazing area 30 m2
12 m2
Orientation (Front/rear) North/South South/North
Building completion date November 2009 May 2012
External wall construction Structural insulated panel (SIP) system
containing rigid foam insulation.
Design U-value: 0.15 W/m2K
Masonry with fully-filled mineral wool
cavity. Design U-value: 0.26 W/m2K
Floor construction Suspended beam and block concrete.
Design U-value: 0.16 W/m2K
Suspended beam and block concrete
floor. Design U-value: 0.23 W/m2K
Roof construction SIP warm roof. Design U-value: 0.15
W/m2K
Cold roof. Design U-value: 0.11
W/m2K
Glazing Double. Centre pane U-value: 0.93
W/m2K
Double. Centre pane U-value: 1.40
W/m2K
Thermal Mass Parameter Medium Medium
(SAP 2009)
Predicted heat loss 120.1 W/K 92.6 W/K (B) 90.1 W/K (C)
Measured air permeability 3.43 m3/h.m
2 9.86 m
3/h.m
2 (B) 9.05 m
3/h.m
2 (C)
Table 4: Summary of the estimations of the HLC of Dwelling A obtained from the three test scenarios
Test scenario Analysis period HLC (W/K) Standard error
(W/K)
Variation from
baseline
LeedsMet
(baseline)
05/12/12 -
10/01/13
133.7 1.9 n/a
Integrated (fans) 20/01/13 –
13/02/13
134.8 2.3 +0.8%
Integrated 27/10/12 –
03/12/12
117.1 2.1 -12.4%
Table 5: Summary of the estimations of the HLC of Dwelling B obtained from the three test scenarios
Test scenario Analysis period HLC (W/K) Standard error
(W/K)
Variation from
baseline
LeedsMet (baseline) 21/12/12 –
10/01/13
114.5 2.3 n/a
Integrated (fans) 09/03/12 –
24/03/12
104.5 2.8 -8.7%
Integrated 21/02/13 –
07/03/13
105.2 1.5 -8.1%
Table 6: Summary of control Dwelling C HLC estimations obtained using LeedsMet coheating
Table 7: Reduction in the thermal transmittance of the external walls of the test dwellings from the baseline test
to the integrated test with fans
Dwelling Baseline test external
wall in situ U-value
(W/m2K)
Integrated (fans) test
external wall in situ U-
value (W/m2K)
External
wall area
(m2)
Reduction in
external wall heat
loss (W/K)
B (test) 0.33 (± 0.01) 0.28 (± 0.01) 84.28 3.9
C (control) 0.34 (± 0.02) 0.30 (± 0.01) 84.28 3.1
Test scenario Analysis period HLC (W/K) Standard error
(W/K)
Variation from
baseline
Baseline
21/12/12 –
10/01/13
105.5 2.2 n/a
Integrated (fans) 09/03/12 –
24/03/12
99.2 2.5 -6.0%
Integrated 21/02/13 –
07/03/13
102.9 1.0 -2.5%
Figure captions
Fig. 1: Test dwelling A (left) and test dwellings B and C (right)
Fig. 2: Linear regression estimation of Dwelling A HLC from each coheating test scenario
Fig. 3: Linear regression estimation of Dwelling B HLC from each coheating test scenario
Fig. 4: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) during the
baseline HLC estimation period
Fig. 5: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) obtained
during the integrated test with fans
Fig. 6: Correlation between 24 hour HLC estimates of Dwelling B (test) and Dwelling C (control) obtained
during the integrated test
Fig. 7: Box plot of the mean hourly internal temperature in Dwelling A during the three test scenarios
Fig. 8: Box plot of the mean hourly internal temperature in Dwelling B (test) and Dwelling C (control) during
the three test scenarios
List of Figure Captions
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Figure(s)Click here to download high resolution image
Highlights
Whole house heat loss test method using dwelling’s own heating system is proposed.
HLC estimates from integrated coheating are compared with electric coheating.
Good agreement between the HLC estimates obtained using each method.
Integrated HLC estimate could be more indicative of how a dwelling performs in-use.
Efficiency of the dwelling’s heating system can also be obtained using test method.
*Highlights (for review)